Pull requests for Datasets

I have a question about the datasets pull requests and how it works?

I can get how the pr in models or spaces works since they’re code but how the pr works for dataset?
And if i wanna make one should i work locally or what ?

It would be really helpful if you can answer me these questions since they’re crucial for a project I’m working on that is based on community contributions.

CC : @severo

If the dataset files you want to edit are in text format, like CSV or JSON, you can edit them directly in the web editor, like you do on models and space. Click “contribute” and edit the file to open a PR. See prompts.csv · fka/awesome-chatgpt-prompts at main for example.

As an alternative, you can download the file, work on it locally, and then go to the parent directory (for example, fka/awesome-chatgpt-prompts at main) and click + ContributeUpload file to open a PR that will replace the file.

Capture d’écran 2023-10-06 à 16.36.17

Finally, you can also do it by cloning the repository and working with git as explained in Pull requests and Discussions.

1 Like